1 / 40

RecMax : Exploiting Recommender Systems for Fun and Profit

RecMax : Exploiting Recommender Systems for Fun and Profit. Amit Goyal Laks V. S. Lakshmanan. University of British Columbia. http://cs.ubc.ca/~goyal. Recommender Systems. Products. Movies. Music. Videos. News. Websites. RecMax – Recommendation Maximization.

marla
Download Presentation

RecMax : Exploiting Recommender Systems for Fun and Profit

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. RecMax: Exploiting Recommender Systems for Fun and Profit AmitGoyal Laks V. S. Lakshmanan University of British Columbia http://cs.ubc.ca/~goyal

  2. Recommender Systems Products Movies Music Videos News Websites

  3. RecMax – Recommendation Maximization • Previous research mostly focused on improving accuracy of recommendations. • In this paper, we propose a novel problem RecMax (short for Recommendation Maximization). • Can we launch a targeted marketing campaign over an existing operational Recommender System?

  4. Consider an item in a Recommender System Flow of information • Because of these ratings, the item may be recommended to some other users. • Some users rate the item • (seed users) RecMax: Can we strategically select the seed users?

  5. RecMax 5 Flow of information • Users to whom the item is recommended • Seed Users Select kseed users such that ifthey provide high ratings to a new product, thenthe number of other users to whom the product is recommended (hit score) by the underlying recommender system algorithm is maximum.

  6. RecMax – Problem Formulation Recommendation List for user v rating threshold of user v (denoted by θv) Number of recommendations are l For a new item i, if expected rating R(v,i) >θv, then the new item is recommended to v The goal of RecMax is to find a seed set S such that hit score f(S) is maximized.

  7. Benefits of RecMax • Targeted marketing in Recommender Systems • Marketers can effectively advertise new products on a Recommender System platform. • Business opportunity to Recommender System platform. • Similar to Influence Maximization problem in spirit. • Beneficial to seed users • They get free/discounted samples of a new product. • Helpful to other users • They receive recommendations of new products – solution to cold start problem.

  8. A key Challenge – Wide diversity of Recommender Systems Similarity functions: Cosine, Pearson, Adjusted Cosine etc Due to this wide diversity, it is very difficult to study RecMax

  9. Outline • What is RecMax? • Does Seeding Help? – Preliminary Experiments. • Theoretical Analysis of RecMax. • Experiments. • Conclusions and Future Work.

  10. Does Seeding Help? • Dataset: Movielens • Recommender System: User-based • Seeds are picked randomly. • Recall that Hit Score is the number of users to whom the product is recommended. Abudget of 500 can get a hit score of 5091 (10x) (User-based)

  11. Does Seeding Help? • Dataset: Movielens • Recommender System: Item-based • Seeds are picked randomly. • Recall that Hit Score is the number of users to whom the product is recommended. Abudget of 20 can get a hit score of 636 (30x) (Item-based)

  12. Outline • What is RecMax? • Does Seeding Help? – Preliminary Experiments. • Theoretical Analysis of RecMax. • Experiments. • Conclusions and Future Work.

  13. Key Theoretical Results • RecMax is NP-hard to solve exactly. • RecMax is NP-hard to approximate within a factor to 1/|V|(1-ε) for any ε> 0. • No reasonable approximation algorithm can be developed. • RecMax is as hard as Maximum Independent Set Problem. • Under both User-based and under Item-based.

  14. Why is RecMax so hard? (1/2) • We introduce a helper problem – Maximum Encirclement Problem • find a set S of size k such that it encircles maximum number of nodes in the graph. A B E C D • Nodes {B,C} encircle node A. • Nodes {B,C,E} encircle node D. • Thus, {B,C,E} encircle A and D.

  15. Why is RecMax so hard? (2/2) • Set {B,C,E} is a solution to Maximum Encirclement Problem (for k=3). • Nodes {A,D} form Maximum Independent Set. A B E • Reduction: Nodes {B,C,E} must rate the new item highly for the item to be recommended to A and D. • RecMax is as hard as Maximum Independent Set, and hence NP-hard to approximate within a factor to 1/|V|(1-ε) C D

  16. Discussion (1/2) • We show hardness for User-based and Item-based methods. • What about Matrix Factorization? • Most likely hardness would remain (future work).

  17. Discussion (2/2) • Since the problem is hard to approximate, does it make sense to study? • YES, as we saw earlier, even a random heuristic fetches impressive gains. • We explore several natural heuristics and compare them. • What about sophisticated heuristics (future work).

  18. Outline • What is RecMax? • Does Seeding Help? – Preliminary Experiments. • Theoretical Analysis of RecMax. • Experiments. • Conclusions and Future Work.

  19. Datasets

  20. Heuristics • Random: Seed set is selected randomly. The process is repeated several times and average is taken. • Most-Active: Top-k users with most number of ratings. • Most-Positive: Top-k users with most positive average ratings. • Most-Critical: Top-k users with most critical average ratings.

  21. Heuristics • Most-Central: Top-k central users.

  22. User-Based Recommender Systems

  23. Comparison – Hit Score achieved • Dataset: Movielens • Recommender System: User-based Most Central, Most Positive and Random perform good here.

  24. Comparison – Hit Score achieved • Dataset: Yahoo! Music • Recommender System: User-based Most Positive, Most Central perform good here.

  25. Comparison – Hit Score achieved • Dataset: Jester Joke • Recommender System: User-based Most Central out-performs all other heuristics.

  26. Key Takeaways • Even the simple heuristics perform well. • With a budget of 300, Most-Central heuristic achieves hit score of 4.4K, 3.4K and 15.6K on Movielens, Yahoo! and Jester respectively. • Depending on the data set, we may encounter a “tipping point” – a minimum seeding is needed for the results to be impressive.

  27. Item-Based Recommender Systems

  28. Comparison – Hit Score achieved • Dataset: Movielens • Recommender System: Item-based Most Central performs good here.

  29. Comparison – Hit Score achieved • Dataset: Yahoo! Music • Recommender System: Item-based Most Central performs good here.

  30. Comparison – Hit Score achieved • Dataset: Jester Joke • Recommender System: Item-based Most Central, Random and Most-Active performs good here.

  31. Key Takeaways • Again, the simple heuristics perform well. • Hit score achieved in Item-based is much lower than in User-based. • Thus, much less seeding is required to achieve maximum possible hit score. • Overall, Most-Central performs well. • The difference of Most-Central with baseline Random is not much. • We need better heuristics (future work).

  32. User-Based vs Item-Based

  33. User-based vs Item-based • Dataset: Yahoo! Music • Initial rise of hit score is steeper in Item-based. • Hit score saturates much earlier in Item-based. Eventual hit score that can be achieved is much more in User-based.

  34. User-based vs Item-based Seed Sets are different in both methods.

  35. Outline • What is RecMax? • Does Seeding Help? – Preliminary Experiments. • Theoretical Analysis of RecMax. • Experiments. • Conclusions and Future Work.

  36. Our Contributions • The main goal of the paper is to propose and study a novel problem that we call RecMax • Select k seed users such that if they endorse a new product by providing relatively high ratings, the number of users to whom the product is recommended (hit score) is maximum. • We focus on User-based and Item-based recommender systems. • We offer empirical evidence that seeding does help in boosting the number of recommendations

  37. Our Contributions • We perform a thorough theoretical analysis of RecMax. • RecMax is NP-hard to solve exactly. • RecMax is NP-hard to approximate within any reasonable factor. • Given this hardness, we explore several natural heuristics on 3 real world datasets and report our findings. • Even simple heuristics like Most-Central provide impressive gains • This makes RecMax an interesting problem for targeted marketing in recommender systems.

  38. Future Work • RecMax is a new problem and has real applications – our work is just the first work. • Developing better heuristics. • Studying RecMax on more sophisticated recommender systems algorithms • Matrix Factorization.

  39. Thanks and Questions RecMax: Exploiting Recommender Systems for Fun and Profit Amit Goyal LaksV. S. Lakshmanan University of British Columbia University of British Columbia http://cs.ubc.ca/~goyal

  40. Heuristics • Most-Central: Top-k central users.

More Related